Reuben Feinman
Research Scientist at Google DeepMind
Cambridge, Massachusetts, United States
Summary
👤
Senior🎓
Top SchoolReuben Feinman is a research scientist with 11 years of AI experience spanning industry and academia, currently at Google DeepMind after leading deep learning efforts at Common Sense Machines. He holds a PhD in Computational Neuroscience from NYU (Google PhD Fellow) and has collaborated with leaders like Yann LeCun and Ruslan Salakhutdinov on self-supervised vision and malware-detection research. Reuben builds large-scale foundation models and ML infrastructure, applying transformers, diffusion models, deep RL and neuro-symbolic methods to 3D generative AI and interactive design. His work is cited 1,000+ times across premier ML venues, and he’s contributed practical adversarial-attack tooling to the popular CleverHans library, including a novel saliency-map method and tutorial. Based in Cambridge, MA, he blends rigorous probabilistic and information-theoretic foundations with production-grade systems engineering to translate cutting-edge research into deployed capabilities.
11 years of coding experience
5 years of employment as a software developer
Doctor of Philosophy - PhD Neural Science, Doctor of Philosophy - PhD Neural Science at New York UniversityBachelor of Science (Sc.B.) with Honors Applied Mathematics, Bachelor of Science (Sc.B.) with Honors Applied Mathematics at Brown University
Github Skills (7)
machine-learning10
adversarial-machine-learning10
tensorflow10
python10
security9
benchmarking8
benchmark8
Programming languages (7)
Github contributions (5)
cleverhans-lab/cleverhans
Oct 2016 - Mar 2017
An adversarial example library for constructing attacks, building defenses, and benchmarking both
Role in this project:
ML Engineer
ML EngineerContributions:41 commits, 4 PRs, 27 comments in 4 months
Contributions summary:Reuben primarily contributed to the implementation of the Saliency Map method within the CleverHans library, introducing a new adversarial example generation technique. They added the necessary code for the saliency map method, including its core algorithm, and also provided a tutorial script for demonstrating its usage. Furthermore, the user refactored code, updated documentation, and fixed issues related to Fast Gradient Sign Method (FGSM), demonstrating a focus on improving the library's functionality and usability for the generation and evaluation of adversarial examples.
benchmarkingrobustnessadversarial-machine-learningsecurityadversarial
rfeinman/pyBPL
Apr 2018 - Jun 2022
Python implementation of Bayesian Program Learning tools (with PyTorch)
Contributions:643 commits, 1 PR, 69 pushes in 4 years 3 months
pytorchpythonsplinespython-implementationbayesian-inference
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.